One of the types of objects in Praat. A **DataModeler** tries to fit *N* data points (*x*_{i}, *y*_{i}) as lying on a particular kind of model function f(*x*; *p*) that depends on *M* *parameters* *p*. The parameters *p* are determined by minimizing the chi squared function χ^{2}(**p**)=Σ ((*y*_{i} - *f*(*x*_{i}, *p*))/σ_{i})^{2}, where the sum is over all *N* data points.

If the individual uncertainties σ_{i}of the data are not known before hand, the model parameter are determined by assuming that all uncertainties σ_{i} are equal. However an independent assessment of the goodness-of-fit cannot be determined in this case. A good guess for σ^{2} for this case could be σ^{2}=Σ (*y*_{i} - *f*(*x*_{i}, *p*))^{2}.

### Drawing

© David Weenink, 2024